Cardholder localization based on transaction data

A computer-implemented method comprising:

• initiating a determination of one or more possible base locations of a cardholder of a payment card;
• associating transactions of the cardholder with a plurality of merchant customer base location profiles, each of the merchant customer base location profiles characterizing geographic base locations of a plurality of customers with which a merchant has had a transaction;
• deriving from the merchant customer base location profiles the one or more possible cardholder base locations;
• and
• providing the one or more possible base locations;
• wherein the merchant customer base location profiles comprise geographic distributions smoothed in accordance with the equation:
• L(zi|mk)=(1−wk)P(zi)+(wkP(zi|mk));
• where:
• zi represents a ZIP code ‘i’;
• mk represents a merchant ‘k’;
• L(zi|mk) is an estimate of the probability that the base location of a customer of merchant k is ZIP code i;
• wk is computed in accordance with the function wk=C(mk)/(C(mk)+a);
• P(zi) is ab overall probability of a cardholder having a base location of a ZIP code i;
• and
• P(zi|mk) is a measured probability of a cardholder having a base location of a ZIP code i for a merchant k;
• wherein P(zi|mk)=C(zi|mk)/C(mk), and C(zi|mk) is the number customers with stated base location ZIP code i for a merchant k or the number transactions by customers with stated base location ZIP code i for a merchant k.

A computer program product, tangibly embodied in a computer-readable medium, the computer program product being operable to cause data processing apparatus to perform operations comprising:

• receiving a request for one or more possible cardholder base locations;
• and
• determining one or more possible cardholder base locations, the determining comprising:
• generating a list of locations using at least one of a plurality of merchant customer base location profiles, each of the locations being a base location associated with a probability that the base location is one of the cardholder base locations, the probability derived from probabilities of a plurality of customers of merchants with which the cardholder has had a transaction having the base location, and determining the one or more possible cardholder base locations from probabilities associated with the list of locations;
• and
• providing the one or more possible cardholder base locations;
• wherein the merchant customer base location profiles comprise geographic distributions smoothed in accordance with the equation:
• L(zi|mk)=(1−wk)P(zi)+(wkP(zi|mk));
• where:
• zi represents a ZIP code ‘i’;
• mk represents a merchant ‘k’;
• L(zi|mk) is an estimate of the probability that the base location of a customer of merchant k is ZIP code i;
• wk is computed in accordance with the function wk=C(mk)/(C(mk)+a);
• P(zi) is ab overall probability of a cardholder having a base location of a ZIP code i;
• and
• P(zi|mk) is a measured probability of a cardholder having a base location of a ZIP code i for a merchant k;
• wherein P(zi|mk)=C(zi|mk)/C(mk), and C(zi|mk) is the number customers with stated base location ZIP code i for a merchant k or the number transactions by customers with stated base location ZIP code i for a merchant k.